Entity

Time filter

Source Type

Wusong, China

Zhou S.,Hefei University of Technology | Zhou S.,Tongling College | Mao M.,Hefei University of Technology | Su J.,Hefei University of Technology
Taiyangneng Xuebao/Acta Energiae Solaris Sinica | Year: 2012

The multi-resolution analysis of wavelet and the good generalization performance of support vector machine were considered sufficiently in the paper, and a wind speed forecasting model based on wavelet and support vector machine was established. The original wind speed sequences are decomposed into coarse components and detail components firstly. Then every wavelet components are separately forecasted with corresponding support vector machine models. Finally, the forecasting results of original wind speed series are achieved by using wavelet reconstruction. The simulation results prove that this method is capable of improving the lag of forecasting values and reducing the error of upheaval point, thereby, it can increase generalization performance and forecasting precision.


Wen C.,Tongling College
2011 International Conference on Electrical and Control Engineering, ICECE 2011 - Proceedings | Year: 2011

The generation of frequent itemsets is the key of association rules mining. Based on bit vectors and its intersection operation of the DLG ideas, this paper presents a new k-frequent itemsets generation algorithm based on bit matrix. The algorithm scans the database only once, using bit matrix of alternative association graph to store, constructing screening conditions to reduce the validation of candidate frequent itemsets in long patterns of frequent itemsets generated effectively. Compared with DLG, experimental results show the effectiveness and accuracy of this algorithm. © 2011 IEEE.


Yongjiang D.,Tongling College
Advanced Materials Research | Year: 2013

Heuristic clustering method based on local information is introduced, and then the label propagation method based on local information is summarized, and the problem of the iterative process and using a random strategy to select a node belongs to the cluster structure are analyzed. Label propagation algorithm base on the similarity of node attributes is improved. At last, the experiments are used to help to discover the efficient and availability of the algorithm, and put the algorithm into preliminary application. © (2013) Trans Tech Publications, Switzerland.


Wen C.,Tongling College
Cybernetics and Information Technologies | Year: 2015

Check-in service, being one of the most popular services in Mobile Social Network Services (MSNS), has serious personal privacy leakage threats. In this paper check-in sequences of pseudonym users were buffered, and a bit matrix for buffered check-in sequences was built, which can achieve privacy guarantee of k-anonymity. The method guarantees that the number of lost check-in locations is minimized while satisfying users' privacy requirements. Besides, it also reduces the cost of finding a trajectory k-anonymity set. At last, the results of a set of comparative experiments with (k, δ)-anonymity on real world datasets show the method accuracy and efficiency.


Wen C.,Tongling College
International Journal of Security and its Applications | Year: 2015

The traditional location privacy protection means mainly employs reliable central server framework, and it mainly applies the anonymous region meeting the k-anonymity at the anonymous server to replace the real location of users. However, the disadvantages of the central server, such as being attacked easily, high communication cost, etc. are disclosed accordingly. In addition, the anonymous method of most locations is oriented for the European style, and it is not applicable to the road network. In order to solve the above problem, virtual user group-based centerless server framework is proposed for solving the location privacy protection in road network. This algorithm mainly forms the virtual user group with several virtual users in several road segments of the road network, and replaces the real position of the user with a certain point in the section, so as to realize the k-anonymity of user position with the increment nearest neighbor query. Since the algorithm adopts the accurate increment nearest neighbor query method, it guarantees the service quality. The experiment proves that the algorithm can reduce the communication cost effectively and improve the application safety. © 2015 SERSC.

Discover hidden collaborations